An MCP (Model Context Protocol) server that adds cosplay characters and personality to LLM responses.
Communication Mode: Stdio only
Just say cosplay <somebody's name>
Install the published package:
npm install -g mcp-cosplayFor local development:
git clone git@github.com:unclemate/mcp-cosplay.git
cd mcp-cosplay
npm install
npm run buildAdd to your MCP client configuration:
Add to your Claude Desktop configuration file (claude_desktop_config.json):
{
"mcpServers": {
"cosplay": {
"command": "node",
"args": ["/path/to/mcp-cosplay/dist/index.js"],
"env": {}
}
}
}Add to your Claude Code configuration file:
{
"mcpServers": {
"cosplay": {
"command": "npx",
"args": ["-y", "mcp-cosplay@latest"]
}
}
}Or if you have installed it globally:
{
"mcpServers": {
"cosplay": {
"command": "mcp-cosplay"
}
}
}Configuration file locations:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
{
"mcpServers": {
"cosplay": {
"command": "npx",
"args": ["tsx", "/path/to/mcp-cosplay/src/index.ts"],
"env": {}
}
}
}Note: Use
tsxto run TypeScript files directly.tsconly compiles TS to JS but cannot execute them.
Note: This MCP server only supports stdio mode communication.
Using the published npm package:
{
"mcpServers": {
"cosplay": {
"command": "npx",
"args": [
"-y",
"mcp-cosplay@latest"
]
}
}
}Or if installed globally:
{
"mcpServers": {
"cosplay": {
"command": "mcp-cosplay"
}
}
}Add cosplay character and personality to text
Parameters:
text(required): Text to cosplaycharacter(optional): Character type (enthusiastic,sarcastic,professional, or custom characters)intensity(optional): Intensity level (0-5, default: 3)context(optional): Context for better character analysis
Example:
{
"text": "Your code looks great!",
"character": "enthusiastic",
"intensity": 4
}Add cosplay character and personality to text (returns processed text only)
Parameters:
text(required): Text to cosplaycharacter(optional): Character type (enthusiastic,sarcastic,professional, or custom characters)intensity(optional): Intensity level (0-5, default: 3)context(optional): Context for better character analysis
Get available character types
Get all available character profiles including custom characters
Get detailed profile for a specific character
Parameters:
name(required): Character name
Add a new custom character profile
Parameters:
name(required): Character namedescription(required): Character descriptionpersonality(required): Personality configurationexamples(optional): Example phrasescategory(optional): Character category
Remove a custom character profile
Parameters:
name(required): Character name to remove
Search characters by name, description, or category
Parameters:
query(required): Search query
Get characters filtered by category
Parameters:
category(required): Category to filter by
Dynamically generate character personality using LLM
Parameters:
characterName(required): Name of the character to generatedescription(optional): Description of the charactercontext(optional): Context for character generationintensity(optional): Intensity level (0-5)examples(optional): Example phrases
Get current configuration
Update configuration
Parameters:
defaultPersonality: Default personality typedefaultIntensity: Default intensity level
Check content safety for violations
Parameters:
text(required): Text to check for safety violations
Get current content safety configuration
Update content safety configuration
Parameters:
enabled: Enable/disable content safety checkscheckMethod: Method to use for content checking (llmorkeyword)confidenceThreshold: Confidence threshold for violation detectionstrictMode: Enable strict mode for content checking
The MCP Cosplay server acts as a "character translator" or "personality filter". It receives standard LLM-generated responses, then performs secondary processing based on your configured rules or character parameters, adding character-specific traits, tone words, and personality before returning the processed text.
- Built-in Characters: Pre-configured personalities (
enthusiastic,sarcastic,professional) - Custom Characters: Add and manage your own character profiles with detailed personality traits
- Dynamic Generation: Generate character personalities using LLM based on descriptions
- Character Search: Search and filter characters by name, description, or category
- Configurable Safety Checks: Enable/disable content safety monitoring
- Multiple Detection Methods: Choose between LLM-based or keyword-based detection
- Adjustable Sensitivity: Set confidence thresholds and strict mode options
- Real-time Monitoring: Check content for safety violations before processing
- Emotion Analysis: Analyze emotional tone and context of input text
- Personality Application: Apply character-specific traits, speech patterns, and tone
- Intensity Control: Adjust personality strength from subtle (0) to intense (5)
- Context Awareness: Consider conversation context for better character portrayal
- Multiple Dialects: Support for different linguistic styles and regional variations
- Cultural Adaptation: Adapt character personalities to different cultural contexts
- Language Style Matching: Maintain consistency across different language styles
The server includes built-in content safety mechanisms that cannot be disabled. These features ensure responsible usage:
- Configurable Safety Levels: Adjust sensitivity thresholds and detection methods
- Context-aware Processing: Automatic adaptation based on conversation context
- Professional Standards: Maintains appropriate language in formal contexts
- Real-time Monitoring: Continuous safety checks during processing
Every response requires additional processing by the MCP server (sentiment analysis -> personalization rewriting), which will increase system response time. You need to optimize your code to ensure efficient processing flow.
The designed personality should be consistent. If a character is an "angry guy", all their responses should conform to this setting, rather than being hot and cold.
See LICENSE file.